Verteon AI

Verteon — Your AI coding agent for VS Code, built to understand, execute, and improve your code while keeping you in control.

Comment

Inspiration

We wanted to make AI-assisted coding feel less like handing control to a black box and more like working alongside a capable developer. Existing AI coding tools can generate code, but developers still need to understand what the AI is doing and remain in control of actions that affect their projects.

Verteon was created to bring an AI coding agent directly into VS Code while keeping transparency, local AI, and user control at the center.

What it does

Verteon is an AI coding agent for VS Code that can understand a project, inspect files, work with the terminal, edit code, run development tools, and interact with Git.

It supports a multi-provider architecture, allowing developers to use local AI through Ollama or connect supported cloud AI providers.

Verteon also includes a safety layer that classifies tool actions and can require user approval for consequential operations while blocking explicitly dangerous commands.

How we built it

Verteon was built as a VS Code extension with a modular architecture.

The system separates the AI provider layer from the agent engine, allowing different models and providers to be used without rebuilding the entire application.

The agent communicates with tools for:

  • File and workspace inspection
  • Code editing
  • Terminal execution
  • Testing and building
  • Git operations
  • Project context and memory

A dedicated safety layer evaluates tool actions before execution. The interface provides streaming responses, live tool activity, approval prompts, provider selection, and workspace information directly inside VS Code.

Ollama provides a local-first option, while the multi-provider architecture allows additional AI providers to be configured.

Challenges we ran into

One of the biggest challenges was giving an AI agent real capabilities without giving it unrestricted control over a developer's machine.

We had to design boundaries around terminal commands, file modifications, Git operations, and other potentially consequential actions.

Another challenge was creating a provider architecture that could support different AI services while keeping the agent engine independent from any particular model.

We also focused on making the interface communicate what the agent is doing in real time instead of hiding tool execution behind a simple chat interface.

Accomplishments that we're proud of

We're proud that Verteon evolved beyond a basic AI chatbot into a working agent architecture capable of interacting with a real development environment.

Key accomplishments include:

  • A functional VS Code AI agent
  • Ollama/local AI support
  • Multi-provider architecture
  • Streaming AI responses
  • Real filesystem and workspace tools
  • Terminal integration
  • Git integration
  • Project inspection and context management
  • Tool risk classification
  • Approval workflows
  • Blocked dangerous-command handling
  • Persistent project memory
  • A dedicated VS Code sidebar interface
  • Packaged VS Code extension
  • Public download/onboarding experience

Most importantly, Verteon is designed around the idea that AI should assist developers without taking away their control.

What we learned

We learned that building an AI agent is much more than connecting an LLM to a chat interface.

The difficult part is building the systems around the model: context management, tool execution, error handling, provider abstraction, permissions, security, and a UI that makes the agent's actions understandable.

We also learned that local AI can be an important option for developers who want more control over where their code and prompts are processed.

What's next for Verteon AI

Our next goal is to continue expanding Verteon's multi-provider ecosystem while improving the agent's reliability and developer experience.

Planned improvements include:

  • More AI providers and models
  • Smarter provider fallback
  • Better project context
  • Improved code-change review
  • More powerful testing and debugging workflows
  • Expanded Git capabilities
  • Better agent memory
  • More granular tool permissions
  • Improved onboarding and installation
  • More polished developer analytics

The long-term vision is for Verteon to become a transparent, controllable AI development partner inside VS Code—one that can handle increasingly complex development tasks while keeping the developer firmly in control.

Built With

Share this project:

Updates

Submission history